星雀优化算法（NOA）matlab代码.zip

%_________________________________________________________________________% % Nutcracker Optimization Algorithm (NOA) source codes demo 1.0 % % % % Developed in MATLAB R2019A % % % % Author and programmer: Reda Mohamed (E-mail: [email protected]) & Mohamed Abdel-Basset (E-mail: [email protected]) % % % % Main paper: Abdel-Basset, M., Mohamed, R. % % Nutcracker optimizer, % % Knowledge-Based Systems, in press, % % DOI: https://doi.org/10.1016/j.knosys.2022.110248 % % % %_________________________________________________________________________% % The Nutcracker Optimization Algorithm function [Best_score,Best_NC,Convergence_curve]=NOA(SearchAgents_no,Max_iter,lb,ub,dim,fobj) %%-------------------Definitions--------------------------%% Best_NC=zeros(1,dim); % A vector to include the best-so-far Nutcracker(Solution) Best_score=inf; % A Scalar variable to include the best-so-far score LFit=inf*(ones(SearchAgents_no,1)); % A vector to include the local-best position for each Nutcracker RP=zeros(2,dim); %% 2-D matrix to include two reference points of each Nutcracker Convergence_curve=zeros(1,Max_iter); %%-------------------Controlling parameters--------------------------%% Alpha=0.05; %% The percent of attempts at avoiding local optima Pa2=0.2; %% The probability of exchanging between the cache-search stage and the recovery stage Prb=0.2; % The percentage of exploration other regions within the search space. %%---------------Initialization----------------------%% Positions=initialization(SearchAgents_no,dim,ub,lb); %Initialize the positions of search agents Lbest=Positions; %% Set the local best for each Nutcracker as its current position at the beginning. t=0; %% Function evaluation counter %%---------------------Evaluation-----------------------%% for i=1:SearchAgents_no NC_Fit(i)=fobj(Positions(i,:)); LFit(i)=NC_Fit(i); %% Set the local best score for the ith Nutcracker as its current score. % Update the best-so-far solution if NC_Fit(i)<Best_score % Change this to > for maximization problem Best_score=NC_Fit(i); % Update the best-so-far score Best_NC=Positions(i,:); % Update te best-so-far solution end end while t<Max_iter RL=0.05*levy(SearchAgents_no,dim,1.5); %Levy random number vector l=rand*(1-t/Max_iter); % Parameter in Eq. (3) %%% Parameter in Eq. (11) if rand<rand a=(t/Max_iter)^(2*1/t); else a=(1-(t/Max_iter))^(2*(t/Max_iter)); end if rand<rand %%%%Foraging and storage strategy mo= mean(Positions); for i=1:SearchAgents_no %%%%% Update the parameter mu according to Eq. (2)%%%%% if rand<rand mu=rand; elseif rand<rand mu=(randn); else mu=(RL(1,1)); end cv=randi(SearchAgents_no);%% An index selected randomly between 1 and SearchAgents_no cv1=randi(SearchAgents_no); %% An index selected randomly between 1 and SearchAgents_no Pa1=((Max_iter-t)/Max_iter); if rand<Pa1 %% Exploration phase 1 cv2=randi(SearchAgents_no); r2=rand; for j=1:size(Positions,2) if t<Max_iter/2 if rand>rand Positions(i,j)=(mo(j))+RL(i,j)*(Positions(cv,j)-Positions(cv1,j))+mu*(rand<5)*(r2*r2*ub-lb); % Eq. (1) end else if rand>rand Positions(i,j)=Positions(cv2,j)+mu*(Positions(cv,j)-Positions(cv1,j))+mu*(rand<Alpha)*(r2*r2*ub-lb); % Eq. (1) end end end else %% Exploitation phase 1 mu=rand; if rand<rand r1=rand; for j=1:size(Positions,2) Positions(i,j)=((Positions(i,j)))+mu*abs(RL(i,j))*(Best_NC(j)-Positions(i,j))+(r1)*(Positions(cv,j)-Positions(cv1,j)); % Eq. (3) end elseif rand<rand for j=1:size(Positions,2) if rand>rand Positions(i,j)=Best_NC(j)+mu*(Positions(cv,j)-Positions(cv1,j)); % Eq. (3) end end else for j=1:size(Positions,2) Positions(i,j)=(Best_NC(j)*abs(l)); % Eq. (3) end end end %%%%%%Return the search agents that exceed the search space's bounds if rand<rand for j=1:size(Positions,2) if Positions(i,j)>ub Positions(i,j)=lb+rand*(ub-lb); elseif Positions(i,j)<lb Positions(i,j)=lb+rand*(ub-lb); end end else Positions(i,:) = min(max(Positions(i,:),lb),ub); end %% Evaluation NC_Fit(i)=fobj(Positions(i,:)); % Update the local best according to Eq. (20) if NC_Fit(i)<LFit(i) % Change this to > for maximization problem LFit(i)=NC_Fit(i); % Update the local best fitness Lbest(i,:)=Positions(i,:); % Update the local best position else NC_Fit(i)=LFit(i); Positions(i,:)=Lbest(i,:); end % Update the best-so-far solution if NC_Fit(i)<Best_score % Change this to > for maximization problem Best_score=NC_Fit(i); % Update best-so-far fitness Best_NC=Positions(i,:); % Update best-so-far position end t=t+1; if t>Max_iter break; end Convergence_curve(t)=Best_score; end else %% Cache-search and Recovery strategy %%% Compute the reference points for each Nutcraker%%%% for i=1:SearchAgents_no ang=pi*rand; cv=randi(SearchAgents_no); cv1=randi(SearchAgents_no); for j=1:size(Positions,2) for j1=1:2 if j1==1 %% Compute the first reference point for the ith Nutcraker using Eq. (9) if ang~=pi/2 RP(j1,j)=Positions(i,j)+ (a*cos(ang)*(Positions(cv,j)-Positions(cv1,j))); else RP(j1,j)=Positions(i,j)+ a*cos(ang)*(Positions(cv,j)-Positions(cv1,j))+a*RP(randi(2),j) end else %% Compute the second reference point for the ith Nutcraker using Eq. (10) if ang~=pi/2 RP(j1,j)=Positions(i,j)+ (a*cos(ang)*((ub-lb)+lb))*(rand<Prb); else RP(j1,j)=Positions(i,j)+ (a*cos(ang)*((ub-lb)*rand+lb)+a*RP(randi(2),j))*(rand<Prb) end end end end %%% Return the reference points that exceed the boundary of %%% search space %% if rand<rand for j=1:size(Positions,2) if